ECE598: Interplay between Control and Machine Learning

Course Information

  • Office Hours: Tu/Th 3-4pm, 145 CSL

  • Lectures: Tu/Th 12:30-1:50pm, Room 3020 ECEB

  • For a complete syllabus, see here.

Course Description

Advanced graduate course focuses on interplay between control and machine learning. The first half of the course focuses on tailoring control tools to study algorithms in large-scale machine learning. In the second half of the course, students will study how to combine reinforcement learning and model-based control methods for control design problems.

We will cover some (or all) of the following topics: empirical risk minimization; first-order methods for large-scale machine learning; stochastic optimization; dissipation inequality; jump system theory; Lur'e-Postkinov type Lyapunov functions; integral quadratic constraints; KYP Lemma; graphical interpretations for optimization methods; adaptive control and ADAM; stable manifold theorem; Lyapunov measure; implicit bias of gradient descent on least square and logistic regression; robust control theory; algorithmic stability; policy gradient on linear quadratic regulator (LQR) problems; learning model predictive control for iterative tasks; zeroth-order optimization and evolutionary strategies; robust control via DK-iteration and IQC-synthesis; adversarial reinforcement learning; imitation learning.

Required Materials

There is no required textbook for the class. All course material will be presented in class and/or provided online as notes. Links for relevant papers will be listed in the resourse section of the course website.

Prerequisites

Math 415, ECE 313, ECE 490 (or any similar course on optimization), and ECE 515 are required. ECE 534 is recommended, but not required.

Grading

60% regular homework sets (3 sets of homework in total); 40% written research report (detailed guidelines for the final projects will be posted in the resource section).